Bess beetles (Odontotaenius disjunctus)survive harsh winters by the process of supercooling. Pinto and Greenberger (unpublished) have demonstrated that the supercooling points (SCPs) of starved O. disjunctus are lower than that of fed beetles. Research has also shown that insects' concentrations of antifreezes increase in cold conditions (Lee et al., 1987). This study tested the effects of both feeding and seasonality on the SCPs of Bess beetles. We froze fed and starved Bess beetles in order to find their SCPs, repeating the same procedure from February through May 2000 so as to examine seasonal trends in their freezing points. We incorporated the data of Pinto and Greenberger from November 1999 to further analyze SCPs with regard to seasonality. NMR analyses of multiple beetle gut samples also enabled us to empirically examine the presence of specific antifreezing agents throughout the season.
We found that there was no significant difference overall in the SCPs of fed vs. starved beetles. There was no seasonal trend in the SCPs within the fed group alone, while the SCPs of unfed beetles increased significantly between November and May (p=0.002) as well as February and May (p=0.011). These data suggest that feeding is the primary cause of elevated SCPs, but when beetles do not feed, other factors such as seasonal fluctuations in antifreeze levels play a large role in regulating their SCPs. In accordance with this theory, NMR analyses showed seasonal variations in the presence of antifreeze. Both glycerol and erythritol were present in samples from February beetles, but by March, gut samples of starved beetles contained neither compound. These results suggested that colligative antifreezes benefit Bess beetles both by lowering SCPs during cold conditions and by serving as a carbohydrate energy source when ambient temperatures increase. Collectively, our results support the notion that a two-part process involving the evacuation of gut contents and the accumulation of antifreeze compounds enable Bess beetles to maximize their supercooling capacity.
Three species of Bess beetles (Family Passalidae) naturally occur in the United States. The eastern species (Odontotaenius disjunctus) experiences harsh winters during which ambient temperatures can drop as low as -10 degrees C. Both larval and adult Bess beetles can survive winter despite the sparse protection provided by rotting tree stumps that house beetle communities because they have evolved the ability to supercool, or suppress their freezing points.
Insects supercool through a two-part process. First, insects evacuate their guts to remove endogenous nucleating agents (Leather et al., 1993). A previous study found that if the beach weevil emptied its digestive tract, it could supercool to -22 degrees C in the winter, whereas insects that had resumed feeding froze at -16.8 degrees C (Bale, 1980). Low supercooling points (SCPs) have been correlated with reduced gut content in moth larvae and beetles (Salt, 1953; Pinto and Greenburg, 1999). Insects accumulate cryoprotecting sugars and proteins to suppress further their SCPs as another part of the supercooling process. Sugars act as colligative antifreezes to increase the solute concentration of hemolymph (Cannon and Block, 1988). Thermal hysteresis proteins (THPs) physically inhibit the growth of ice crystals.
Although cryoprotectants are conducive to supercooling and thus survival, insects must exert a large amount of energy to synthesize them. In order to elevate hemolymph concetrations, insects must produce and conserve glycerol and other colligative antifreezes instead of using their precursors for energy (Lee, 1991). THPs are large, bulky molecules made of many repeating amino acid units (Graham et al., 1997), and their production and maintenance ultimately decreases the growth and reproductive resources of an insect. Since these molecules are so energetically expensive, supercooling insects would benefit from producing and retaining them in a temperature-dependent manner. The flesh fly, elm leaf beetle and milkweed bug synthesize antifreezes so that hemolymph concentrations reach 2.4 times that of normal levels when insects are exposed to 0 degrees C (Lee et al., 1987). Initial findings suggest that hormones and temperature-sensitive enzymes regulate the production of cryoprotectants in other species (Danks, 1996).
The supercooling ability of many insects fluctuates seasonally. Mountain pine beetles collected in autumn supercool to temperatures ranging from -10 to -20 degrees C. The supercooling points of insects collected from the same colonies in winter fell to between 25 and -35 degrees C. When insects were tested in the spring, their supercooling points had risen to that of fall levels (Bentz and Mullins, 1999). Furthermore, physiologists have correlated the seasonality of supercooling points with colligative antifreeze concentrations. Olson and colleagues (1998) suggested that the fire-colored beetle larvae were more resistant to innoculative freezing during the winter, a process partially facilitated by the seasonal modification of colligative antifreeze concentrations. A study of the Alpine Weta demonstrated that their supercooling points decreased in winter, which corresponded with elevated hemolymph osmolarity (Raml°v et al., 1992).
This study examined the influences of food intake and seasonality on the supercooling points of Bess beetles. Cryoprotectant levels were investigated as possible factors contributing to seasonal differences in insect freezing points. We hypothesized the following: (1) beetle SCPs would be lowest in the winter and increase throughout the spring, (2) starved beetles would have lower SCPs than fed beetles overall, and (3) colligative antifreeze concentrations would peak in the winter and decrease throughout the spring. To test our hypotheses, we froze fed and starved O. disjunctus to find their SCPs and took nuclear magnetic resonance spectra (NMR) to identify and track the presence of colligative antifreezes over a period of four months. Bess beetles were chosen as subjects because they are large and easy to work with, do not appear in the scientific literature and are known to supercool (Pinto and Greenberger, unpublished).
Materials and Methods
Experiments were performed between February and May 2000. Bess beetles were ordered from the Carolina Biological Supply Company (Burlington, NC). Upon arrival, beetles were weighed and separated into two groups with comparable mean weights and variance. Both groups were housed at room temperature in covered cages in a dim room and were misted with aged water daily for six to eight days before supercooling. One group (fed) was supplied with decaying oak pulp and unbleached paper towels, and a second group (starved) was supplied with only unbleached paper towels to prevent dessication. Beetles were weighed immediately before supercooling to determine weight loss.
Twelve beetles were supercooled in the February group (six fed and six starved), and twenty-four beetles (twelve fed and twelve starved) in the March and May groups (Table 1).
Table 1. Experimental design for supercooling Bess Beetles. Twelve beetles - six fed and six starved - were supercooled in November and in March, and twenty-four beetles - twelve fed and twelve starved - in April and May. * denotes experiment performed by Christobel Pinto and Sarah Greenberger, Swarthmore College, fall 1999.
After supercooling, February beetles were allowed to recover with food and water and then frozen and dissected, while March and May beetles were kept frozen until their guts were removed for NMR analyses.
Three fed and three starved beetles were supercooled at a time. A 4-gauge copper-constantan thermocouple was wrapped around each beetle with Saran Wrap between the prothorax and mesothorax, and attached at the other end to an Omega Microprocessor thermometer (type J-K-T Thermocouple, model HH23) to measure body temperature. Each beetle was then suspended individually in a glass jar to avoid inoculative freezing (Figure 1). The stoppered jars were placed together in a styrofoam cooler containing 700 mL 95% ethanol to reduce the cooling rate, and the entire unit was placed in a -20 degree C freezer.
Figure 1: Apparatus used for supercooling Bess beetles. A 4-gauge copper-constantan thermocouple was wrapped around each beetle using Saran Wrap between the prothorax and mesothorax. The beetle was then suspended in a glass bottle which was placed in a foam cooler containing 95% ethanol and then into a -20 degree C freezer..
Temperatures were recorded from each of the six thermocouples every ten minutes until they dropped to 0 degrees C, after which they were recorded every minute until at least two minutes after all the beetles exhibited an exotherm.
Nuclear Magnetic Resonance (NMR) Spectroscopy Analysis
13C nuclear magnetic resonance (NMR) spectra were obtained from gut samples following supercooling experiments in order to identify colligative antifreezes in the beetles. Spectra were collected from February beetles, fed March beetles and starved March beetles. February beetles could not be separated into fed and starved treatments because the groups were stored together following supercooling.
NMR samples consisted of homogenized guts from four beetles spiked with approximately 2 mL D2O. A micro-spatula and a long-necked disposable pipette were used to pack the sample into the bottom of a 0.5 cm o.d. NMR tube.
Fourier transformed carbon spectra were obtained at 100 Mhz on a Brunker DRX-400 NMR spectrometer under conditions of complete hydrogen decoupling at 20 degrees C. The sample was not spun, and data were collected following 8,400 scans (approximately 10 hrs.). Spectra were manipulated using a line broadening of 10Hz.
13C NMR spectra were compared by aligning the peaks around 130 and 175 ppm corresponding to oleic and palmitoleic acids found in cell membranes (Buchanan and Storey, 1983). The following candidate antifreezes were evaluated for their presence in the beetle guts by comparing gut spectra to pure chemical spectra available from Aldrich (Milwalkee, WI): erythritol, ethylene glycol diethyl ether, D-fructose, alpha-D-glucose, glycerol, inositol, D-mannitol and sorbitol. Final antifreeze identification was determined by adding 50 ÁL of an approximately 0.30M antifreeze solution suspended in D2O to the homogenized guts and re-acquiring the 13C NMR scan. If peaks that were previously present grew, we inferred that the added antifreeze was present naturally in the Bess beetles.
The supercooling points of Bess beetles were analyzed according to month using a one-way ANOVA, split by fed/starved. These tests allowed us to examine seasonal SCP relationships within each group, but due to the added splitting procedure, only p-values 0.025 and below were considered significant.
There was no significant difference overall between the mean supercooling point (SCP) of fed beetles and starved beetles (Figure 2). When each group was examined separately, starved beetles showed a significant seasonal trend in SCP (F=4.792, p=0.008, DF=3). Mean SCP of starved beetles in May was -5.3 degrees C, which was significantly higher than that of starved beetles in November (-6.7 degrees C; p=0.002) and February (-6.4 degrees C; p=0.011), as shown by pairwise comparisons. Fed beetles showed no significant seasonal difference in SCP (F=0.869, p=0.469, DF=3).
Figure 2. Mean supercooling points and standard errors for groups of fed and starved Odontaenius disjunctus from November 1999 to May 2000. Sample size is indicated by numbers above and below each point.
Nuclear Magnetic Resonance (NMR) Spectroscopy Analysis
The 13C NMR spectra of beetle gut samples consistently contained peaks at 130, 175 and 185 ppm corresponding to ketones in oleic and palmitoleic acids. A number of large peaks in the shielded single bond region (40-10 ppm) also appeared in all spectra, with some variation between acquisitions. When the spectra were aligned using the ketone peaks as references, these single bond carbon peaks generally overlapped. Comparison between spectra required no more than a 3 ppm shift. Since the colligative antifreezes we investigated had peaks between 60 and 100 ppm, we concentrated on that region of the NMR spectrum.
Each of the four February spectra contained a broad peak generally spanning from just below 70 ppm to just below 80 ppm. A sharper peak appeared in all spectra between 60 and 65 ppm. These results excluded D-fructose and D-glucose as possible antifreezes because those sugars produce peaks around 105 ppm. Inositol was also eliminated as a possible candidate because results did not contain a peak in the upper 60's. To identify the antifreezes, we spiked each of the four February samples with one of the following sugars: erythritol, ethylene glycol diethyl ether, glycerol or sorbitol. While mannitol's pure spectrum was consistent with that of the beetle gut, we lacked sufficient gut material to spike a fifth sample. Adding ethylene glycol diethyl ether and sorbitol to February samples caused novel peaks to emerge in the spectrum (data not shown). Since these peaks did not appear in the original sample, those spectra suggest that these two antifreezes were not produced at detectable levels by the February beetles.
Comparisons between original and spiked sample spectra suggest the presence of erythritol and glycerol in February gut samples. Two peaks, one at 62 ppm and the other at 72 ppm, grew sharper and taller when erythritol was added to a February sample (Figure 3). The published spectrum of pure erythritol contains two peaks at 65 and 75 ppm. When glycerol was added to the February sample, two peaks at 63 ppm and 73 ppm grew sharper and taller (Figure 4). The reported 13C NMR spectrum of pure glycerol contains two peaks at 64 and 74 ppm. Both sets of peaks probably shifted upfield due to a changing the environment from D2O for the pure chemical spectrum to beetle gut material for the spiked spectrum. The 10 ppm distance between the two peaks in both sets remained constant between published and experimental results.
Figure 3. NMR spectra from February beetle guts (black) and the same sample spiked with erythritol (blue). The spectra were aligned using the ketone peak at 132 ppm as a reference. Two peaks in the plain gut sample, at 62 and 72 ppm, grew sharper in the erythritol-spiked sample.
Figure 4. NMR spectra from February beetle guts (black) and the same sample spiked with glycerol (red). The spectra were aligned using the ketone peak at 131 ppm as a reference. Two peaks in the plain gut sample, at 63 and 73 ppm, grew sharper in the glycerol-spiked sample.
When erythritol and glycerol were added simultaneously to the fed March sample, peaks at 68.5 and 78.5 ppm grew and peaks at 72 and 82 ppm emerged. We determined that the 72 and 82 ppm peaks corresponded to erythritol based on the 10 ppm distance between peaks, position downfield and relative heights of the peaks. Erythritol carbons should be less deshielded than glycerol carbons because they are surrounded by an additional oxygen. The two peaks were of equal heights and therefore conform to the chemical structure of erythritol, which contains two atoms of each distinct carbon type. Since these peaks were not apparent in the unspiked March fed sample, we concluded that fed March beetles contained no detectable levels of erythritol.
We determined that the 68.5 and 78.5 peaks corresponded to glycerol based on the same criteria. Glycerol carbons should be more deshielded than erythritol carbons because glycerol contains one less alcohol group. Since glycerol has two identical carbons and one distinct carbon, one of the peaks should be half the height of the other. The shorter peak, corresponding to the central carbon, should be more deshielded than the taller peak, corresponding to the terminal carbons, because the central carbon is flanked by twice the number of alcohol functional groups. The observed peaks matched this theoretical description. Since the glycerol peaks were present in the original March fed sample and grew upon addition of the polyol, we concluded that fed March beetles contained detectable levels of glycerol (Figure 5).
When erythritol and glycerol were added simultaneously to the starved March sample, novel peaks at 68.5 and 78.5 as well as 72 and 82 ppm emerged. Since these peaks did not appear in the unspiked sample, we concluded that starved March beetles did not contain detectable levels of erythritol or glycerol (Figure 5).
Figure 5. The NMR spectrum of fed March beetles (red) contained peaks at 68.5 ppm and 78.5 ppm, corresponding to glycerol. The NMR spectrum of starved March beetles (black) did not contain these peaks. Erythritol peaks were absent from both spectra.
The results do not support the hypothesis that fed beetles consistently freeze at higher temperatures than starved beetles; however the two groups demonstrated clear seasonal differences in their supercooling abilities. Starved beetles tended to freeze at lower temperatures than fed beetles in the winter, though the difference diminished with the arrival of spring and completely disappeared by early May (Figure 2). Changes in cryoprotectant concentrations may have contributed to this trend.
Environmental triggers give rise to a general pattern of seasonal changes in cryoprotectants. In freeze-tolerant larvae of the goldenrod gall fly, antifreeze concentrations usually begin to increase in the fall, reach a maximum concentration, then decrease by late spring. Changes in colligative antifreeze concentrations result in fluctuating supercooling capacities (Leather et al. 1993). February and March NMR spectra suggest that antifreeze levels in O. disjunctus decrease similarly between winter and spring. Nevertheless, we only observed a significant SCP increase in the starved beetles. Somme (1982, 1985) argues that gut content is the primary factor influencing supercooling capabilities. Our results also suggest that fed beetles benefit less from seasonal antifreeze production because potential ice-nucleating agents in their guts override any additional freeze suppression. Conversely, secondary factors, such as cryoprotectants, may determine insect SCPs when the emptied digestive tract does not contain ice-nucleating food particles.
NMR analyses detected glycerol and erythritol in Bess beetle gut samples, which strengthens other findings that most insects synthesize only one or two types of antifreeze (Storey and Storey, 1991). Multiple cryoprotectants may benefit insects in a number of ways. The synthesis of different antifreeze compounds allows an organism to respond to separate environmental cues, enabling them to maximize efficiency and synchronize precisely with the onset of winter. A drop in temperature below a certain threshold could trigger the production of one antifreeze; the insect could synthesize another compound in response to late autumn food shortages, a mechanism that would protect against early frosts. Minimizing antifreeze production prior to freezing temperatures conserves energy. In addition, each antifreeze offers its own benefits such as accessible fuel storage and metabolic rate reduction. For instance, glycerol increases cold hardiness without retarding enzymatic activity. Arthropods using this colligative antifreeze often synthesize an additional cryoprotectant to slow metabolism and conserve energy (Zachariassen, 1985). Third, synthesizing multiple cryoprotectants may reduce possible toxic effects associated with high concentrations of a single compound (Leather et al., 1993).
Not surprisingly, our NMR analyses indicated that O. disjunctus synthesize glycerol, the most common colligative antifreeze. Somme (1982) screened 96 terrestrial arthropod species for polyols and sugars, finding 64 that produced high levels of glycerol. He concluded that most insects with extensive supercooling capacity utilize glycerol. The compound probably facilitates supercooling by stabilizing enzymes at low temperatures and protecting them from denaturation (Storey and Storey, 1991).
Selective glycerol metabolism might explain differences in the presence of colligative antifreezes between fed and starved March beetles. Both beetle treatments already reabsorbed erythritol by this point in the spring. Neither group needed the antifreeze in their warming environment, and they probably slowed or stopped production of both antifreezes. Fed March beetles may have ingested enough nutrients to prevent them from breaking down glycerol as an energy source. However, starved beetles may have utilized glycerol in the absence of food. NMR analyses could not quantify glycerol hemolymph concentrations, though March beetles probably possess less antifreeze than February beetles. Had the glycerol levels of the February beetles been as low as those of March, February starved beetles should have metabolized all of their glycerol as well.
Temperature sensitivity of antifreeze production in Bess beetles should be investigated to begin deciphering the seasonality of antifreeze accumulation and potential control mechanisms. Future experiments could test the effects of keeping the harvested beetles at a temperature much lower than 25 degrees C prior to freezing, which would more accurately represent the winter and early spring natural environment of O. disjunctus. Researchers should be warned that housing beetles at 4 degrees C reduces their metabolism so that starved beetles do not appear to lose significant amounts of mass. If winter temperatures trigger cryoprotectant synthesis, one would expect to find greater antifreeze levels in beetles housed in the cold. The possibility that Bess beetles can synthesize colligative antifreezes rapidly confounded preliminary results testing this hypothesis. Because beetles were sacrificed by freezing, they most likely manufactured antifreezes prior to death. While it was assumed that increased cold-hardening takes place over weeks to months, laboratory experiments have revealed that increasing the production of cryoprotectants may occur within a two-hour exposure to low temperatures (Chen et al., 1987). Therefore, experiments which examine seasonal trends in accumulation of cryoprotectants are likely to be more successful if an alternative method of beetle sacrifice is utilized. Despite this restriction, however, we were able to distinguish seasonal trends in the synthesis of glycerol and erythritol; perhaps the mechanisms regulating their production are mediated by the termination of diapause rather than by immediate temperature fluctuations. Further experiments may also investigate the relationship of diapause to cold-tolerance in this species; these studies may provide further insight into the coordination of gut evacuation and the accumulation of antifreeze compounds in an effort to maximize supercooling.
We would like to thank Sara Hiebert for her persevering mentorship and insights through this experiment. Thanks to Jocylene Noveral for teaching us how to supercool insects, Tom Valente for his entimology knowledge, and Collin Purrington for his statistics suave. We would also like to thank Ahamindra Jain and Jon Huber for their NMR expertise and patience with our lack thereof.